Green versus Grey Framing: Exploring the Mechanism behind the Negative Footprint Illusion in Environmental Sustainability Assessments

Abstract

Given the complexity of assessing the environmental sustainability of products, consumers rely on cognitive strategies to simplify complex information and develop quick judgments, often referred to as heuristics, when processing eco-information. One of these heuristics is called ‘the Negative Footprint Illusion’: Consumers erroneously estimate the total environmental impact of a combination of a green and non-green product as lower than the same non-green product alone. In this research, we test this bias and explore its underlying mechanism. We evoke a more summative vs. more evaluative mindset by framing the response scales negatively (in terms of environmental damage, referred to as ‘grey scaling’) vs. positively (in terms of environmental friendliness, referred to as ‘green scaling’). This is carried out by using an online between-subject experiment in which respondents either respond on an evaluative response scale (green scaling), or a summative response scale (grey scaling). A hamburger and bio-apple were used as stimuli (either shown together or apart). First, the results show that the negative footprint is only apparent in the green scaling condition. Second, respondents who score higher on environmental concern show a stronger negative footprint illusion for the green scaling condition. Our study not only elucidates the cognitive mechanisms driving the negative footprint illusion but also offers strategic directions for both theoretical advancement and practical applications in environmental decision-making, highlighting effective ways to mitigate this bias.

1. Introduction

The need for action on tackling climate change has reached a make-or-break moment; without immediate intervention, we will reach an increase of three degrees Celsius of the world’s average surface temperature this century, causing detrimental effects on economies and human health [1]. A key factor in reducing greenhouse gas emissions is in the hands of consumers and their efforts to shift everyday consumer choices towards more environmentally sustainable options [1,2]. Moreover, the inclusion of ‘responsible consumption and production’ as the 12th goal in the UN 2030 agenda for sustainable development underscores the pivotal role of consumers in advancing towards a more sustainable world. However, although the availability and consumption of sustainable products has been growing [3,4], the adoption of sustainable practices in society is lagging behind [5]. One of the causes of this problem relates to the fact that consumers have a hard time distinguishing sustainable from unsustainable options, as this requires an advanced level of analytical skills and knowledge [6]. Given the complexity of this information, consumers rely on simplified eco-information schemes (e.g., labels, eco-ratings, eco-metrics) to determine the sustainability of products [7]. Although these eco-information schemes make the information easier to grasp, they can also lead to less desirable outcomes.

Research has shown that consumers, when faced with a combination of green and non-green products, wrongly estimate the total environmental footprint of this combination to be lower than the footprint of the same non-green product alone, a bias referred to as ‘the Negative Footprint Illusion’ [8]; for a review, see [9]. For example, when consumers are asked to estimate the environmental impact of a burger (non-green) and an eco-labeled salad (green), they will erroneously combine the footprint of both products into an average instead of adding up the environmental impact of both the burger and the salad. As a result, consumers estimate a lower environmental footprint for the combination of the burger and the eco-labeled salad than for the burger alone. This effect has been replicated in several contexts other than food choices, like conventional versus energy-efficient buildings [10,11] and regular versus hybrid cars [12], which seems to be insensitive to the quantity of the added non-green items [13] and prevails in within- as well as in between-subject designs [10], all providing strong support for the robustness of the illusion. Erroneously and consistently underestimating the total environmental impact of product choices is problematic, as this could lead consumers to increase consumption to reduce their footprint, with overconsumption as a result. Understanding the underlying mechanism of this bias is crucial, as such insights can be instrumental in devising effective strategies to mitigate its impact.

Researchers are still studying the mechanism behind this bias [8,10,14,15], but have suggested that an averaging bias might underpin the Negative Footprint Illusion. According to this, consumers’ natural response to this type of information would be to evaluate the products in terms of vices vs. virtues, or good vs. bad. This evaluative mindset leads people to average the total environmental impact instead of adding up the impact of the different components, resulting in the Negative Footprint Illusion. Therefore, inducing a more summative mindset should correct for this and should eliminate the illusion. One way of testing this is by priming consumers with a summative mindset by means of other tasks that demand additive processing, prior to estimating the environmental footprint of combinations of products [15]. Research by Holmgren et al. (2021) shows that inducing a summative mindset can indeed temper the Negative Footprint Illusion [15].

Another way of studying the mechanism is by using different rating scales to manipulate the way participants/consumers conceptualize environmental impact. The impact of response scale characteristics on respondents’ perception and response behavior is well-documented [16]. An essential consideration in this context is the presence of an underlying objective reference point on the scale. For instance, incorporating an objective reference point such as a frequency per week (e.g., once a week, twice a week) provides respondents with a clear and tangible basis for interpretation. This labeling ensures that respondents can easily align their responses with the appropriate position on the scale. However, the descriptors on a scale can also be tied indirectly to an underlying objective framework, e.g., a scale assessing environmental impact (ranging from not impactful to very impactful). In this scenario, respondents might engage in a process of attempting to objectively evaluate the extent of the impact, and subsequently translate this assessment into the appropriate numerical representation on the scale. This might lead to a summative, more arithmetic mindset. In contrast, a scale featuring descriptors like “environmental friendliness” lacks a clear objective underlying scale. The term “friendliness” inherently embodies a subjective perception. Consequently, scales labeled with such subjective descriptors may elicit a more evaluative process, wherein respondents rely on personal judgment and subjective interpretation to assess and assign values. As such, the way a scale is presented may have a significant influence on the results obtained and is inherently part of the measurement process.

Similarly, the possible influence of the response scale format has already been investigated in the article in which the illusion is introduced [8]. More specifically, in study 2, the authors investigated whether the illusion is moderated by the rating scale format that participants use to evaluate the environmental impact (the footprint) of the menus, by comparing a color-coded rating scale of impact to a quantitatively anchored rating scale [8]. However, the expected interaction effect of the type of scale—in which the more evaluative scale would strengthen the negative footprint illusion compared to the quantitative scale—was not found. One possible explanation of this insignificant effect could be that the manipulation of the scales was suboptimal at inducing evaluative vs. additive processing. Both scales were framed in a negative way that we further refer to as ‘grey’, as the ecological feature of products is expressed in terms of damage to the environment: very low to very high environmental impact on a 7-point scale, either depicted by color-coded footprints or by grey footprints indicating CO2 emissions. A positively framed scale, further referred to as ‘green’ however, would be a better way to induce an evaluative mindset, as this expresses the ecological attribute in terms of environmental friendliness, which is more evaluative in nature. So far, all studies on the Negative Footprint Illusion have used grey-framed rating scales to measure consumers’ evaluation of the sustainability of choice options, e.g., [8,10,11,13,14,15]. Hence, it is imperative to close this gap by exploring the conditions under which the negative footprint illusion manifests. Additionally, it is essential to delve into how interindividual variables may shape this illusion. Notably, the influence of environmental concern on the negative footprint illusion remains an important aspect that requires further investigation.

In this study, we test the effect of a green vs. a grey-framed rating scale on the negative footprint illusion and hypothesize that while adding a green to a non-green food product necessarily increases total environmental impact (footprint), consumers will erroneously estimate the total environmental impact of the combination of the green and non-green product lower than the same non-green product alone, and expect that this effect will be stronger in the green scale condition (less to more environmentally friendly).

The project’s findings provide valuable input in defining policy and strategies to encourage environmentally sustainable consumption, which can help consumers in accurately interpreting eco-information. Furthermore, it may give us information on how mindset inducement may influence the interpretation of eco-labels.

2. Materials and Methods

We ran a two (organic apple present vs. absent) by two (green vs. grey-framed response scale) between-subjects experiment. The presence of an organic apple vs. its absence operationalizes the idea of adding a presumable sustainable product to a non-sustainable product. All participants were presented with a picture of a meal that consisted of a burger with or without an apple. These stimuli were taken from study 2 in the paper that introduced the Negative Footprint Illusion [8].

Participants then rated on a seven-point scale their response to the question: “According to you, how would you rate the environmental impact of the meal above?”. In the grey vs. green framing conditions, the low and high end of the rating scale were labeled ‘very damaging for the environment—not damaging for the environment’ vs. ‘not environmentally friendly—very environmentally friendly’, respectively. Next, participants completed two multi-item scales using five-point agreement scales (with response categories ‘strongly disagree, somewhat disagree, neither agree nor disagree, somewhat agree, strongly agree’). The GREEN scale probes environmental concern [17]. To control for acquiescence response style, we used the balanced version of the GREEN scale, which consists of three regularly worded and three reverse-worded items [18]. To assess socially desirable responding, we used the Impression Management (IM) scale, which consists of five regularly worded and five reverse-worded items [19]. An instructed response item was also included among the IM items to control for respondent carelessness (i.e., For this question, please select ‘strongly disagree’) [20]. The questionnaire concluded with background data (year of birth and gender) and the ‘useme’ question as a quality check (“In your honest opinion, should we use your data in our analyses in this study? Please answer honestly, this will not affect your compensation”) [20].

3. Sample and Results

3.1. Sample

We collected data on Prolific. From the initial N = 402 participants, in line with our preregistration (https://aspredicted.org/7NK_LXQ, accessed on 1 January 2024) and methodological recommendations [20], three were removed for failing the instructed response item, and three more were removed for negatively responding to the ‘useme’ question, leaving an effective sample of N = 396. In this sample, the age ranges from 18 to 93 (M = 43.78, SD = 14.02), and 42.68% are female.

3.2. Results

In the analysis, we focus on participants’ rating of the environmental friendliness of the menu as the key dependent variable, henceforth labeled ‘eco-perception’. The hypotheses pertain to the effect of two experimental factors (presence vs. absence of an eco-component, i.e., the organic apple; and use of a grey vs. green-framed rating scale). We also have data for two covariates: Environmental Concern (EC) and Impression Management (IM). Because the covariates are latent variables and in line with recent methodological recommendations concerning experiment data-analysis, we analyze the data using structural equation modeling [21] in Mplus 8.4 [22]. For reasons of transparency, unconditional ANOVA results are provided in Appendix A.

3.2.1. Preliminary Analysis

In a preliminary confirmatory factor analysis, we model EC and IM as two freely correlating latent constructs. As both scales are balanced (i.e., they contain equal numbers of regularly and reverse-worded items) and to control for acquiescent responding, we construct balanced item parcels for both scales; please refer to [23] for a detailed explanation about why and how balanced parceling works. For EC, we construct three item parcels by taking the average score of the first regular item with the first reversed item, the second regular item with the second reversed item, etc. For IM, we use the same approach. This results in three balanced item parcels for EC and five balanced item parcels for IM. The resulting CFA model shows an acceptable fit to the data (chi²(19) = 56.771, p < 0.001, RMSEA = 0.071, CFI = 0.970, TLI = 0.956, SRMR = 0.037), and the model estimates provide evidence in support of the scales’ internal consistency, with composite reliability estimates exceeding the common cutoff value of >0.70: CREC = 0.856 (95% CI = [0.862, 0.903], CRIM = 0.825 (95% CI = [0.833, 0.885]). EC and IM show a small but statistically significant correlation of r = 0.193 (95% CI = [0.082, 0.305]). We will therefore control for IM in subsequent analyses.

3.2.2. Main Analyses

For the main analysis that aims to test the hypotheses, eco-perception serves as the dependent variable, while the presence versus absence of the apple (coded 1 vs. 0) serves as the independent variable. We use rating scale framing (green vs. grey) as the grouping variable so that the model estimates can be readily compared across the grey vs. green-framed rating scale formats. EC and IM are included as covariates (i.e., they are freely correlating antecedents of the dependent variable). Finally, to account for heterogeneity in the experimental effect, we specify an interaction term between the experimental dummy (0 = apple absent, 1 = apple present) and environmental concern, using the Latent Moderated Structures approach [24,25], implemented as the ‘xwith’ procedure in Mplus.

Figure 1 depicts the operational model (with key estimates).

Table 1. Regression coefficients (with eco-perception as the dependent variable).

	Green Rating Scale			Grey Rating Scale			Difference
	B	LL	UL	B	LL	UL	Est.	LL	UL
Eco-component	0.512	0.108	0.915 *	−0.024	−0.440	0.392	0.535	−0.043	1.114
EC × Eco-component	0.433	−0.040	0.906	0.121	−0.331	0.574	0.311	−0.342	0.964
Environmental concern (EC)	−0.582	−0.988	−0.177 *	−0.493	−0.794	−0.191 *	−0.090	−0.591	0.412
Impression Management (IM)	0.166	−0.064	0.397	0.310	0.058	0.563 *	−0.144	−0.483	0.195

Note: B = Unstandardized regression coefficients with eco-perception (rated on a seven-point scale) as the dependent variable. LL/UL = lower/upper limit of the 95% confidence interval. Eco-component is coded 0 = organic apple absent, 1 = organic apple present. Environmental concern and impression management are standardized latent factors (with mean zero and standard deviation one). EC × Eco-component is the latent interaction between environmental concern and the eco-component dummy (estimated using the xwith procedure in Mplus). * = statistically significant at p < 0.05.

reports structural regression coefficients with their related 95% confidence intervals per group (green-framed rating scale condition vs. grey-framed rating scale condition) and the difference between these estimates.

Note that the negative footprint illusion results in a positive effect here, given that higher scores on the rating scale indicate higher environmental friendliness and thus a lower footprint. In the green frame condition, adding the organic apple has a significantly positive effect on eco-perception (unstandardized B = 0.512, 95% CI = [0.108, 0.915]), whereas in the grey frame condition, adding the organic apple has close to no effect (unstandardized B = −0.024, 95% CI = [−0.440, 0.392]). The difference between these two effects, however, is not statistically significant (Est. = 0.535, 95% CI = [−0.043, 1.114]). Environmental concern has a significant main effect on eco-perceptions, such that people who are more environmentally concerned display lower eco-perceptions in both the green and grey frame conditions. Eco-perceptions are positively related to impression management, even though this relation is statistically significant only in the grey frame condition (Unstandardized B = 0.310, 95% CI = [0.058, 0.563]) and not the green frame condition (Unstandardized B = 0.166, 95% CI = [−0.064, 0.397]); the difference in regression coefficients is not statistically significant (Est. = −0.144, 95% CI = [−0.483, 0.195]). Finally, the results do not show statistical significance (at p < 0.05) for the moderation effect of environmental concern on the experimental effect (presence of the organic apple). However, the interaction plots (see Figure 2) suggest that the effect of adding an eco-component to the unsustainable meal affects eco-perceptions only in the green frame condition for people who score average to above-average on environmental concern.

4. Discussion

In the current study, we argue that consumers erroneously estimate the total environmental impact of a combination of the green and non-green product as lower than the same non-green product alone (i.e., the Negative Footprint Illusion), and that evaluating the eco-attribute of products on a rating scale using a green frame (i.e., in terms of eco-friendliness instead of damage) will lead consumers to use a more evaluative mindset in which consuming green products is desirable and by extension good for the environment, which would translate into a stronger negative footprint illusion. Our results provide partial support for our hypotheses and show that the Negative Footprint Illusion is prevalent among consumers with higher levels of environmental concern. Furthermore, we show that among these consumers, adding an eco-component to the unsustainable meal affects eco-perceptions in the green frame condition.

Our results are in line with research of Holmgren et al. (2021), in which priming consumers with a summative mindset prior to estimating the environmental footprint of combinations of products reduced the Negative Footprint Illusion [15]. In their study, participants had to conduct another task which required summative reasoning, prior to the critical judgment task (i.e., estimating the total environmental footprint of the combination of non-green and green products). The researchers suggest that the illusion can be tempered when people are nudged towards a summative mindset. However, even in the summative mindset condition, participants were still not able to make correct estimations of the environmental footprint of the combination of green and non-green items. A correct judgement would result in an increase of the total environmental impact, while the responses in their experiments did not significantly differ from zero (not in a positive nor a negative direction), pointing towards a zero-footprint illusion.

The fact that consumers who are highly concerned about the environment are more prone to this bias might seem somewhat surprising, as one would expect that these consumers are better informed and would therefore be able to make a better judgment about the true environmental impact of (combinations of) products. However, these findings align with the well-established theory on self-regulatory focus which states that people rely on one of two primary self-regulatory strategies to attain their goals: promotion versus prevention [26]. People with a promotion focus are motivated to obtain gains and to maximize positive outcomes, while people with a prevention focus are driven by the desire to avoid losses and to minimize negative outcomes. When the context aligns with the primary self-regulatory focus—i.e., self-regulatory fit—people experience higher motivation and satisfaction [26]. Applied to the current context, framing response scales in a positive way highlights positive outcomes and gains. Adding an eco-friendly component to the unsustainable meal aligns with this green frame and leads to a positive impact on eco-perceptions. This might especially be the case for consumers with higher levels of environmental concern, as their motivation to obtain gains related to preserving nature might be more outspoken; therefore, the positive effect of adding the eco-component might be exaggerated. The match between consumers’ environmental concern, the green frame, and the additional green component might therefore strengthen the Negative Footprint Illusion. However, when response scales use a grey frame, the focus lays on preventing losses and avoiding negative outcomes. Introducing the green product does not align with the current primary strategy (i.e., prevention), and this mismatch between framing and the eco-component might diminish the latter’s positive effect. Again, for consumers who are highly concerned about the environment, this mismatch might have a stronger effect on eco-perceptions than for consumers with lower levels of environmental concern.

Research on the influence of individual differences in environmental concern and beliefs is still scarce. One of the few studies on individual differences on the Negative Footprint Illusion suggests that estimating the environmental impact of a combination of products is influenced by people’s compensatory green beliefs [27]. That is, people who believe that everyone has a certain budget of environmental damage that they can ‘spend’ and compensate for by showing green behavior, seem to be more susceptible to the Negative Footprint Illusion than people who do not (or to a lesser degree) hold these beliefs [27,28]. Since compensatory green beliefs are negatively correlated with environmental concern, this is somewhat contradictory to our findings [28]. However, in an attempt to further explore individual variation in susceptibility to the illusion, Threadgold et al. (2022) were not able to replicate the effect of compensatory beliefs [29]. Given these mixed results, future research should further explore the effect of individual differences in green concern and beliefs.

The present findings also provide valuable methodological insights into the possible effects of response scale formats. The framing of a rating scale as either green or grey appears to have an important influence on research outcomes. Traditionally, studies on the Negative Footprint Illusion have predominantly employed what we have labeled ‘grey’ rating scales (where the ecological feature of products is expressed in terms of damage to the environment), consistently yielding evidence for the illusion. However, our study deviates from this trend. Contrary to expectations, the grey-framed rating scale exhibits no effect, while the green-framed rating scale shows noteworthy results. This discrepancy highlights that research outcomes are not independent of the chosen rating scale. Rather, the scale actively contributes to the measurement process as it may bring respondents in a specific mindset. These results emphasize that experimental manipulations are not the sole determinants of outcomes; the choice of the rating scale also plays a crucial role.

The outcomes of this study offer practical insights for policymaking. First, the findings indicate that individuals with a strong environmental concern are not necessarily impervious to biased thinking about the environment; in fact, heightened environmental concern may even amplify such effects. Consequently, policymakers need to recognize that those with elevated environmental concerns should also be a focal point of targeted interventions and that influencing environmental concern might not necessarily lead to increased pro-environmental behavior. Second, the results suggest that the use of grey framing in communication about the ecological sustainability of products may be beneficial, as it is likely to reduce the negative footprint illusion (compared to the use of green framing). By understanding and leveraging the impact of different framings of environmental sustainability, policymakers can devise interventions that address biased thinking among consumers.

One notable limitation of this study lies in the exclusive utilization of a specific set of stimuli, namely the burger and bio-apple. Consequently, the generalizability of the findings to a broader spectrum of stimuli remains uncertain. However, it is important to note, as outlined in the introduction, that the Negative Footprint Illusion has previously manifested in diverse contexts. This suggests that analogous results could be replicated in alternative contexts. Nevertheless, for robust validation, further investigation across a range of stimuli is imperative.

In the preregistration, we also posited a competing hypothesis in which grey framing strengthened loss aversion, but this rationale is unlikely to apply if participants do not express an intention to buy or consume, and simply rate the eco-impact of a menu. We therefore do not discuss this alternative hypothesis in detail, but suggest that it is mainly relevant for future research that focuses on consumption rather than perception.

5. Conclusions

In the current study, we tested whether the Negative Footprint Illusion can be explained by the fact that consumers process eco-attributes in an evaluative way and tested this by manipulating the framing direction of the response scales. Our findings provide initial support for our hypotheses and further suggest that experimental manipulations could have a differential impact on eco-perceptions, depending on individual differences in environmental concern.

The significance of these findings lies in the revelation that the Negative Footprint Illusion is contingent upon specific conditions, particularly the mindset of individuals. This discovery holds dual importance. Firstly, it underscores the potential for policymakers to leverage the observed phenomenon that inducing a summative mindset in humans can mitigate the Negative Footprint Illusion. Secondly, the study illuminates the substantial impact of response scales on experimental outcomes, demonstrated by the absence of the Negative Footprint Illusion in grey scaling as opposed to green scaling. Next, our results show that people who are more environmentally concerned show a higher tendency to fall prey to the Negative Footprint Illusion. This shows that higher environmental concern does not necessarily lead to better environmental outcomes.

These results underscore the need for future research to delve deeper into these processes, seeking additional insights into the influence of environmental concern. Furthermore, future research might replicate this study with other stimuli. By unraveling the intricacies of these factors, we can enhance our understanding and contribute to more nuanced strategies for addressing and mitigating the Negative Footprint Illusion.